SMART Modular Technologies manufactures memory products in a nearly 300,000 square-foot factory in Malaysia. Like a lot of companies that have started down the path toward AI-driven automation, also commonly known as Industry 4.0, we weren’t following a lofty ambition at the outset. We were following a focused, pragmatic instinct to solve a business problem that was threatening to impact our costs, process efficiency and product quality—the pillars of success in our intensely competitive industry.

In many ways, the key point of how we got on the path to Industry 4.0 can be traced to the way our top executives framed a discrete problem with their eyes on the big competitive picture.

Quality Control under a microscope

Our memory components are embedded in systems that face some of the most demanding conditions—from military and aviation to outer space, whereby extreme quality is the critical performance benchmark. While low-defect products rely on a chain of high-quality processes, the stage that matters most is at the end: product inspection. In a traditional process environment, our well trained inspectors examine memory chips under a microscope for cracked solder joints, damaged components or many other kinds of imperfections. But inherent in this human-based inspection model was a very basic pain point that came down to human behavior.

As the labor market in Malaysia has tightened in recent years—due in part to an aging work force—companies in our sector have seen an increase in both labor costs and employee turnover. As we found out first-hand, this dynamic was especially at play in high-stress jobs like chip inspection. What made it especially challenging was the significant amount of training inspectors needed to do their job, a costly investment effectively lost when an inspector moved on.

The first steps of a transformation journey

With this imperative fresh in their minds, a group of our top executives were invited to a supply chain event held by the quality management team from IBM, for which we are a key supplier. While the executives were impressed by the transformation principles outlined by IBM during the event, one factor made their takeaways stand out from others they heard: IBM has extra credibility for the simple reason that IBM had put its own principles into action. We believed that while others told us to transform, IBM had shown it.

The first steps of our journey began when we engaged a team of IBM manufacturing transformation specialists to perform a weeklong comprehensive Industry 4.0 assessment at our factory. What we got back wasn’t a sales pitch but a roadmap and a framework for prioritizing transformation initiatives. Not surprisingly, quality control automation was at the top of the list.

Deep learning meets Quality Control challenges

Working with a team from IBM, we implemented an automated inspection solution that scans memory cards on the production line for defects. To a large extent, its ability to pick out flaws comes from learning what a “good” product looks like and using AI to visually detect deviations. The fact that we have a “high mix/low volume” product line, with many different products produced in small runs, presented a big challenge: how to make training technically and economically viable.

Our answer was to use the deep learning algorithms within IBM Maximo Visual Inspection—in combination with the global IPC610 standard for inspecting printed circuit boards—to teach the system to recognize the good and the bad for each product line. It demonstrated how AI could be flexibly applied to make automation work in complex environments. When one of the AI-powered robots identifies a defective product, it automatically sets it aside into an NG (no good) tray for more detailed inspection by our QC specialists.

From automation to connected manufacturing

Where our company is going next with AI-driven automation cogently illustrates how Industry 4.0 transformation is in many ways a cumulative journey, built on a foundation of real-time data. In our next phase, we’re planning to expand the use of AI to identify failure patterns that can be used to trace the root cause of defects, like the need to make adjustments to production equipment settings or maintenance schedules. That will further set the stage for us to achieve perhaps the fullest realization of our connected manufacturing vision, with equipment up and down the production lines using automated, machine-to-machine (M2M) communications to work out problems on their own.

In the big picture, we believe our Industry 4.0 transformation positions us as a disruptor competitor, rather than the disrupted. As we move closer to having fully digitized, connected manufacturing operations, we expect a quantum improvement in production efficiency, marked by lower costs, improved productivity and fewer production disruptions. From what we’ve seen, we expect our automated inspection processes to increase production yield by 10%, while increasing overall production throughput by 20%.

Was this article helpful?

More from Business transformation

Business process management (BPM) examples

7 min read - Business Process Management (BPM) is a systematic approach to managing and streamlining business processes. BPM is intended to help improve the efficiency of existing processes, with the goal of increasing productivity and overall business performance. BPM is often confused with other seemingly similar initiatives. For example, BPM is smaller in scale than business process reengineering (BPR), which radically overhauls or replaces processes. Conversely, it has a larger scope than task management, which deals with individual tasks, and project management, which…

Using generative AI to accelerate product innovation

3 min read - Generative artificial intelligence (GenAI) can be a powerful tool for driving product innovation, if used in the right ways. We’ll discuss select high-impact product use cases that demonstrate the potential of AI to revolutionize the way we develop, market and deliver products to customers. Stacking strong data management, predictive analytics and GenAI is foundational to taking your product organization to the next level.   1. Addressing customer inquiries with an AI-driven chatbot  ChatGPT distinguished itself as the first publicly accessible GenAI-powered…

Integrating AI into Asset Performance Management: It’s all about the data

3 min read - Imagine a future where artificial intelligence (AI) seamlessly collaborates with existing supply chain solutions, redefining how organizations manage their assets. If you’re currently using traditional AI, advanced analytics, and intelligent automation, aren’t you already getting deep insights into asset performance? Undoubtedly. But what if you could optimize even further? That’s the transformative promise of generative AI, which is beginning to revolutionize business operations in game-changing ways. It may be the solution that finally breaks through dysfunctional silos of business units,…

IBM Newsletters

Get our newsletters and topic updates that deliver the latest thought leadership and insights on emerging trends.
Subscribe now More newsletters